Summary/Abstract: Autism spectrum disorder (ASD) is a heterogeneous disease with an unknown etiology. The global increase in ASD incidence suggests that genetics alone is unlikely to be the major driver of ASD, but that the increased prevalence is likely due to altered exposures to environmental factors. In fact, we know that numerous environmental exposures (nutrients, chemicals, stress, etc.) impact child health, typically exerting their toxicity through either metabolites or perturbations in endogenous pathways, making metabolomics and protein analysis key emerging technologies to elucidate the relationships between these exposures and ASD. But how do we directly measure these early life exposures? Central to our study is the use of novel tooth matrix biomarkers, which take advantage of the incremental manner (similar to tree growth rings) of the developmental biology of teeth. The techniques that we have developed allow us to temporally distinguish exposure between the 2nd trimester, 3rd trimesters, and postnatal periods, enabling identification of the sensitive life stages for biological perturbations in fetal and neonatal development most strongly associated with ASD risk. For the present application, we will perform the first untargeted metabolomics analysis of ASD teeth to delineate unique alterations in corresponding autism and non-autism children. This will be supported by the first targeted highly multiplexed protein analysis of teeth (92 proteins) for delineating biological pathways of interest, including inflammation, oxidative stress and those associated with neurobiological processes. We will use novel statistical methodology, weighted quantile sum regression (WQS) that addresses time-varying effects of high-dimensional mixtures, and increases power when compared to traditional methods to discover biomarkers and biological pathways associated with ASD. Discovery will be performed on 40 ASD case-control sibling pairs, and replication on an independent population of 35 unrelated case-control pairs. Our method is a non-invasive advancement in technology to obtain direct and repeated measures of biomarkers associated with early life etiology of ASD.